Zhiliang Qin

Write Failure Analysis for Bit-Patterned-Media Recording and Its Impact on Read Channel Modeling

In this paper, we present a write failure analysis for bit-patterned media recording (BPMR) with formulas derived to calculate its probability for a given set of system parameters. The write failure model derived in this work includes the effect of the recording head and media characteristics and that of a hard disk spindle motor speed variations. Such a model can help assess the performance of the writing strategy for BPMR. It is also found that since the media local parameters associated with each bit directly affect both the write failure rate and read-back signal, the read-back signal actually contains information regarding whether a bit is correctly recorded. This may be exploited by coding and detection schemes to reduce the impact of write failure on bit-error-rate performance.

Iterative reduced-complexity multiuser detection based on Chase decoding for synchronous turbo-coded CDMA system

The conventional multiuser detector (MUD) based on the a posteriori probability (APP) algorithm has an exponential computational complexity in terms of the number of users. In this paper, we propose a low-complexity iterative multiuser receiver for synchronous turbo-coded code-division multiple-access (CDMA) systems. The proposed receiver is based on the Chase decoding algorithm that was previously used to decode turbo product codes. Simulation results show that the proposed receiver can significantly reduce the computational complexity with slight performance degradation compared with the APP MUD over highly correlated channels. Moreover, in this paper, we develop a numerical approach to analyze the convergence behavior of iteratively decoded CDMA channels based on density evolution technique. Analytical results are presented and shown to provide a reasonable match with what is observed in simulation.

Modeling, detection, and LDPC codes for bit-patterned media recording

In this paper, we present a thorough and comprehensive study for bit-patterned media recording (BPMR), from a signal processing and coding perspective. We first propose a recording-physics-based generic channel model for BPMR, which includes all the major characteristics and impairments of the system. It also provides a fair basis for the performance comparison of different coding and detection schemes. We further propose various channel algorithms and techniques for BPMR, including a two-dimensional (2D) equalization scheme with one-dimensional (1D) generalized partial response (GPR) target to mitigate inter-track interference (ITI) and media noise, a maximum a posteriori (MAP) detector for BPMR with write errors, various low-density parity-check (LDPC) codes, as well as the iterative detection and decoding schemes. The corresponding performance gains are illustrated at 4Tb/in2.

Modeling of 2-D Magnetic Recording and a Comparison of Data Detection Schemes

Two-dimensional magnetic recording (TDMR) together with shingled magnetic recording (SMR) are technologies proposed to extend the life of conventional granular magnetic recording. The grain flipping probability (GFP) model has been proposed to mimic the performance of micromagnetic ( μ-mag) simulations for the purpose of signal reproduction. Other work in TDMR includes the proposal of a Gaussian mixture model (GMM) that produces improved likelihood information at the output of the detector, combined with low density parity check (LDPC) codes. The contribution of this paper is threefold. First, we aim to simulate a TDMR/SMR recording system with the GFP model, both with and without the GMM detector, and with various random and structured LDPC codes, of both 4 k and 16 k block lengths, to determine areal densities that might be achieved. Second, we perform a comparison of the various model order reduced (MOR) GFP implementations to compare the effect of writing with various factors taken out of the picture. Third, we perform a validation of the GFP model and the setup as a whole, by running the system with a parameter set close to that of conventional recording. The results of these experiments give an assurance of the validity of our model, give an indication of the expected density that might be achieved in a TDMR/SMR system, and give a direction for which parameter(s) in magnetic recording systems might be optimized to yield the most gain.

Markov Chain Monte Carlo Based Detection for Two-Dimensional Intersymbol Interference Channels

Recent advances in ultra-high-density data storage technologies have led to two-dimensional (2-D) intersymbol interference (ISI) in the system. The optimum detection method for 2-D ISI channels has prohibitive computational complexity which makes it impractical. In this paper, we propose a Markov chain Monte Carlo (MCMC) based 2-D detection algorithm whose complexity grows polynomially with the ISI size. When turbo decoded in conjunction with a low-density parity check (LDPC) channel code, the performance of the proposed detection method is evaluated for ultra-high-density bit-patterned magnetic recording (BPMR) systems. Its computational complexity is also analyzed. Our study shows that the proposed detection method significantly reduces the computational burden, while achieving performance better than the BCJR based detection method.

Timing and Written-In Errors Characterization for Bit Patterned Media

Synchronous writing in BPMR has been recognized as a crucial yet challenging issue. It has been shown that position jitter and switching field distribution can lead to spatially uncorrelated random write failures. On top of that, the spindle speed variation and other mechanical vibration may lead to accumulative phase drift during writing which causes long streams of insertion/deletion write failures. However no experiments have been conducted to quantitatively testify the later concept, nor provide precise written-in error characteristics when both phenomena are present in BPMR systems. This gap of understanding is filled by this work through spinstand and hard disk measurements and analysis. It is shown that timing inaccuracy not only introduces insertion/deletion write failures but also given rise to substantial increase of substitution (random) write errors. It also reveals that instead of the commonly accused spindle motor speed variation, timing error in BPMR based HDD may well be the result of estimation error due to limited or improperly configured timing preambles.

Channel capacity and soft-decision decoding of LDPC codes for spin-torque transfer magnetic random access memory (STT-MRAM)

Spin-torque transfer magnetic random access memory (STT-MRAM) has emerged as a promising non-volatile memory (NVM) technology, featuring compelling advantages in scalability, speed, endurance, and power consumption. In this paper, we focus on large-capacity stand-alone STT-MRAM, and investigate the channel capacity and the viability of applying low-density parity-check (LDPC) codes with soft-decision decoding to correct the memory cell errors and improve the storage density of STT-MRAM. We propose to use LDPC codes with short codeword lengths, with the reliability-based min-sum (RB-MS) algorithm for decoding. Furthermore, we propose to use the capacity-maximization criterion to design the quantizer and minimize the number of quantization bits. Simulation results demonstrate the potential of applying short-block-length LDPC codes with soft-decision decoding to improve the yield and push the scaling limitation of STT-MRAM.

Turbo Multiuser Detection Based on Local Search Algorithms

The full-complexity soft-input/soft-output (SISO) multiuser detector based on the a posteriori probability (APP) algorithm has a computational complexity growing exponentially with the number of users. In this paper, we consider the multiuser detection problem from a combinatorial optimization viewpoint and develop a novel class of SISO multiuser detectors based on local search (LS) heuristics. By restricting the search to a small subset of the solution space and producing the APP based on a candidate list of potential solutions, the proposed detector can achieve bit-error-rate (BER) performance close to that of the APP algorithm with a significantly lower computational complexity.

Iterative reduced-state decoding for coded partial-response channels

The conventional iterative decoding based on the Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm rises exponentially in terms of channel memory length. In this paper, we propose a low-complexity soft-input/soft-output (SISO) channel detector based on tentative hard estimates fed back from the outer decoder in the previous iteration. The computational complexity of the proposed detector is polynomial in terms of the channel memory length. To demonstrate the performance/complexity tradeoff of the proposed detector, we present simulation results for 9-tap, 11-tap, and 12-tap channels. We show that the proposed detector significantly reduces the computational complexity with only slight performance degradation compared to the full-complexity BCJR algorithm.

A Reduced-Complexity Iterative Receiver Based on Simulated Annealing for Coded Partial-Response Channels

In this paper, we propose a low-complexity soft-output channel detector for turbo equalization over coded partial-response channels based on the simulated annealing (SA) algorithm. It is shown that the proposed detector can significantly reduce the computational complexity with only slight performance degradation compared to the BCJR algorithm